The Influence of Economic and Social Factors on the Choice of Search Procedures Used by Third Party Recruiters

Abstract

Many organizations employ the services of third party recruiters in order to find suitable candidates for their free positions. Yet, we still know little about what drives search agents to use the specific recruitment tools they use. Based on institutional theory, we predicted three economic (costs and time investment, speed and longevity) and three social factors (reactions by client organizations and by candidates, diffusion in the field) to influence search agents' choice of recruitment procedures. Additionally, institutional theory suggested several moderators such as the contract structure between the search agent and their clients, the likelihood of follow-up work, the availability of suitable candidates and the economic uncertainty in the field. Assumptions were tested via an online-survey distributed to 149 search agents working in the Netherlands. Hierarchical linear modelling generally supported the relevance of the economic and social predictors proposed as well as the working of several, albeit not all, moderators proposed. [PUBLICATION ABSTRACT]

Headnote

Abstract

Many organizations employ the services of third party recruiters in order to find suitable candidates for their free positions. Yet, we still know little about what drives search agents to use the specific recruitment tools they use. Based on institutional theory, we predicted three economic (costs and time investment, speed and longevity) and three social factors (reactions by client organizations and by candidates, diffusion in the field) to influence search agents' choice of recruitment procedures. Additionally, institutional theory suggested several moderators such as the contract structure between the search agent and their clients, the likelihood of follow-up work, the availability of suitable candidates and the economic uncertainty in the field. Assumptions were tested via an online-survey distributed to 149 search agents working in the Netherlands. Hierarchical linear modelling generally supported the relevance of the economic and social predictors proposed as well as the working of several, albeit not all, moderators proposed.

Keywords: recruitment, recruitment tools, headhunter, search agent, institutional theory, job advertisement.

Introduction

Although not actively searching for a new job, you sometimes may be tempted to consider changing jobs. For example, when opening a newspaper or surfing the internet, you may find well-looking job openings: Papers display colourful ads and job sites (e.g., monsterboard) show promising possibilities. Other times, you may receive a surprising phone call from someone you have never met before but who seems to know quite a bit about you and who wants to gain your interest for leaving your current organization and for becoming a new employee elsewhere.

After all, employees and the way they work compromise the crucial difference between successful and unsuccessful organisations (Heraty & Morley, 1998) and with increasing technological advancements and shortening product life cycles, the major source of competitive advantage will be the individual worker. This makes attracting and selecting suitable employees a crucial need for any organization that strives to keep a competitive advantage (Schneider, 1987). The projected shortages for top talent will likely increase the competition among organisations over available qualified labor pools (Collins & Han, 2004) and has increased practitioner and academic interest in understanding how companies can focus their recruitment efforts to set them apart from labor market competitors (Collins, 2007).

Several parties can be involved in these recruitment activities. Many organizations search for new employees themselves and may even have a particular department specializing in the recruitment of new employees. Yet, searching for new employees can be a demanding, conflict-laden and time-consuming process - particularly given that the most desirable candidates are often not even looking for a new job at a given moment but are often working successfully and happily in their current jobs in other organizations. Therefore, organizations often decide to outsource the recruitment process by hiring the service of a search firm or 'headhunters', third-party agents who are paid a fee by employers for finding job candidates for them (Finlay & Coverdill, 2002). In this scenario, the organization acts as a client buying the service of the search agent. The search agent in turn searches for suitable candidates and raises their interest in becoming new employees of the clients' organization. Yet, despite the relevance of professional search agents for filling both managerial and specialist functions in organizations, relatively little is known about how these firms go about recruiting candidates for their clients - and why they may favour one approach over another.

The goal of the current study is an explorative review of the different search procedures used by third party recruiters and a theory-grounded explanation as to why recruiters may choose one procedure over another. This way, the paper makes the following three contributions: First, it provides more information about the executive search industry by offering an overview over the procedures currently used as well as over professional recruiters' perceptions about several of these procedures. Second, it provides not only descriptive data but also an explanation as to why and when recruiters may choose to focus on one or the other recruitment approach. Based on institutional theory, we will discuss that not only economic considerations drive the choice of a certain recruitment procedure, yet much more concerns surrounding the social legitimacy of third party recruiters - and with this, the dependencies that recruiters experience on both their clients and the candidates whom they are trying to recruit. Finally, the current study offers another test of institutional theory in an otherwise as of yet understudied domain, thus broadening the generalizability of this theory across different domains. In the following, we will discuss the role of third party recruiters, followed by a discussion of institutional theory, the theoretical backbone of our subsequent hypotheses.

Third Party Recruitment

Searching for and selecting new employees can be a time-consuming and potentially conflict-ridden activity, particularly when candidates are scarce, currently securely situated in other organizations, and when the mutual risk for hiring organizations and candidates is high as the suitability of the candidate for the available position is difficult to evaluate (Khurana, 2002). Thus, organizations in need of scarce employees often solicit the help of a third party recruitment agency, also called "headhunter" or "professional search agent". Third-party recruiters are third-party agents who find job candidates for employers and who are paid by these employers (Finlay & Coverdill, 1999). While reliable data on the frequency with which third-party recruiters are involved in the filling of positions are hard to find, results from the 1990s suggest that one-third or more of firms regularly use such services, with more recent data suggesting these proportions to be stable, if not growing (Finlay & Coverdill, 2002).

Despite their relevance in the recruitment and selection business and the wide media- and business attention that they have received, research on the activities of third party recruiters is scarce. Past research has primarily addressed the reasons why organizations hire third party recruiters in the first place (Finlay & Coverdill, 1999, 2002) and the functions that third party recruiters play in filling structural holes between their client organizations and the candidates (e.g., Coverdill & Finlay, in press; Finlay & Coverdill, 2000; Khurana, 2002). In short, this research shows that headhunters can become more attractive to hiring managers than managers' own in-house HR-department for at least three reasons. First, headhunters are more likely to search in specialized networks of currently employed and well-performing candidates, rather than addressing unemployed, unhappy, and potentially unqualified candidates, thus ensuring a higher base-rate of qualified candidates for the organization to choose from. Second, headhunters build their own specialized databases with potential candidates from particular industries, thus being more likely to have fast and wide access to a great number of possible candidates which in turn should lead to faster and better results. Third, they arguably provide better and more unconditional service to the hiring manager than would be the hiring manager's own HR-department (Finlay & Coverdill, 1999).

This service essentially consists of (a) identifying a broad pool of suitable candidates, oftentimes primarily "hidden" candidates who are otherwise not actively on the market, of (b) narrowing the candidate pool to those who would be interesting for the organization to hire and who, in turn, would accept an offer if provided, and finally of (c) managing the complex employment transaction (Finlay & Coverdill, 1999; Fish & Macklin, 2003; Khurana, 2004). Related to these activities, much research has been undertaken on the scientific recommendations and practical applications of selecting among candidates (i.e., Klehe, 2004; Van Iddekinge & Ployhart, 2008), yet hardly any research can be found on the factors that determine how agents search for potential candidates in the first place.

Some of these search procedures, such as calling potential candidates at their work, sometimes under false pretence, and exploiting their insecurities to lure them away from their current employers and from jobs that candidates were actually quite happy with (Bowe, Bowe, & Streeter, 2000; Finlay & Coverdill, 2002), have given professional search agents a bad or at least questionable name, and even deception and disloyalty are sometimes part of search agents' repertoire in order to get their jobs done (Bowe et al., 2000; Finlay & Coverdill, 2002). Additionally, as 'matchmakers of the labor market', professional search agents capitalize on their own databases and specialized professional networks of possible candidates. Other procedures, however, do not differ all that much from procedures that hiring organizations themselves might undertake when searching for new talents, such as issuing advertisements or - if the hiring organization's HR department had the time for it - sieving through online networks to identify candidates with suitable profiles. Overall, however, it is not clear yet which factors may motivate professional search agents to employ one or the other search procedure available.

Institutional Theory

The current paper tries to address this question with the help of Oliver's (1991) model of strategic responses to institutional processes. The model aims to explain when and why organizations make decisions the way they do and choose certain procedures over others and has been tested and applied to questions such as responsiveness to work-family issues (Goodstein, 1994; Ingram & Simons, 1995), the strategic use of governmental processes (Caeldries, 1996), and the use of different personnel selection procedures (Klehe, 2004; König, Klehe, Berchtold, & Kleinmann, 2010). Based on institutional theory (DiMaggio & Powell, 1983; Zucker, 1977) and resource dependence theory (Pfeffer, 1981; Pfeffer & Salancik, 1978), Oliver (1991) proposed that all organizations attempt to obtain stability and legitimacy. In order to achieve this, organizations need to respond to different institutional pressures. Each of these pressures is associated with a certain cause or intended objective. Causes generally fall into the categories of economic fitness or social fitness (Oliver, 1991; see top of Figure 1). In the case of executive recruiters, economic fitness implies a short-term cost control and a fast income as well as long-term economic benefits arising from finding sustainable solutions for clients. Social fitness refers to the social legitimacy that the recruitment procedure holds in the eyes of the clients, the candidates, and the wider society in terms of the procedure's diffusion in the field. Whether conscious or not, such institutional constituents impose numerous regulations and expectations on organizations, and the more dependent a recruiter is on any constituent and the more control this constituent has over the recruiter, the more effort the recruiter will expend in satisfying this constituent's demands (Oliver, 1991).

When it comes to such control, clients, for example, may decide to switch their providers of recruitment services or may refuse to pay the search agent until a certain benchmark has been reached (e.g., until the candidate signs the contract with the client organization, usually referred to as 'no cure no pay' policy; Finlay & Coverdill, 2002). Candidates may refuse to start conversations or may even refuse the eventual job-offer (Finlay & Coverdill, 2002). Agents themselves will have to watch their time and financial budget when initiating a search, and finally, the industry overall will hold certain standards defined by the procedures perceived to be disseminated in the field (Oliver, 1991).

In a related line, there will be certain conditions under which any of these constituents hold more power than under others. E.g., if a candidate refuses to start negotiations, it may not matter as much if the agent has ten other candidates to choose from. If, however, this candidate is the only suitable candidate in sight, losing the candidate due to an inappropriate recruitment approach may be fatal for the agents' ability to satisfy his or her contract. And finally, one needs to consider the context of recruitment in terms of the procedures that are standard in the field (Oliver, 1991). We will explain each of these factors in more detail below.

Economic Fitness

Organizations strive for economic accountability and rationalization in the form of short as well as long term economic gains (Oliver, 1991). Furthermore, past research on personnel selection found the costs of the procedures to be a major criterion for organisation in deciding whether to use a specific procedure or not (Klehe, 2004; König et al., 2010).

Costs and time investments. Recruitment is no cheap business (e.g., Lavelle, 2003) with professional search agents usually being among the most expensive option available to organizations in need of new employees (Maher, 2003). A prime source of these costs, besides the costs associated with active advertising and networking, is manpower, i.e., the time that the search agent and his or her assistants spend (or estimate to spend) on the search for suitable and available candidates. Often, agents will not know whether these investments will pay off. Agents may or may not get reimbursed for their efforts (particularly those related to the prospected time that they invest into their search), depending on their form of contract with their client and their subsequent success in securing a match.

Different professional search agents usually agree on different types of contracts with their clients, with contracts differing in the degree to which they defend the interests of the client or that of the search agent. Most favourable for clients are usually 'no cure no pay' contracts under which the agent gets paid only if and only when he or she produces a suitable candidate for the client's free position. Here, agents usually get paid once the candidate signs the contract with the client or actually starts working there, but quite regularly, clients also withhold a certain percentage of the fee for the duration of a certain guarantee period (often around six months - i.e., long enough for the candidate to prove his or her worth and to have lost his or her ties to the old employer).

Things look differently under less solution based and more effort based conditions, e.g., when agents get paid from the start for the work that they project to invest into the project or when they work from a retainer. Also, agents regularly negotiate an exclusivity period, a period of time that an agent can search for candidates without the competition of other search agents hired by the same client. When this period (of usually eight weeks) ends without results, the client may give the assignment to another search agent, but until then, the agent acts as the sole recruiter of suitable candidates for this particular job, thus giving him or her a little more time to find a suitable candidate.

Overall, it is quite likely that these different types of contract will moderate the influence that costs and time investment will have on professional search agents' decision to use a specific recruitment procedure. With a 'no cure no pay' contract, agents will invest money and particularly time into their search at their own risk, given that a failure to produce an interesting as well as interested candidate before a competitor will result in wasted effort and thus a loss. The benefits to engage in such investment are thus relatively low. With an effort-based contract, in contrast, investments undertaken will immediately pay offand agents may thus be less hesitant to invest heavily into the search. Thus we propose:

Hypothesis 1a: The more cost- and time-intensive professional search agents perceive a specific recruitment procedure to be, the less likely they are to use this procedure.

Hypothesis 1b: This effect will be moderated by the type of contract that search agents use with their clients. The more their contracts tend towards a client-friendly solution based contract ('no cure no pay'), the less likely they will be to make heavy investments upfront, while the reverse will be true for effort based contracts.

Speed of getting a response. The higher the degree of economic gain perceived to be attainable from conforming to an institutional pressure, the more likely the organization is to conform (Oliver, 1991). For professional search agents, the primary - and oftentimes the sole source of income is to place candidates with clients. Thus, search agents usually have a high incentive to accomplish this as fast as possible and the speed of finding a suitable candidate is also a prime reason why hiring managers employ the services of search agents in the first place (Finlay & Coverdill, 2002). Not only is the target position at the client oftentimes vacant 'right now' and thus in need of a suitable candidate as soon as possible, but search agents often find themselves competing with other agencies when looking for suitable candidates. This will particularly be true if they have not been guaranteed an exclusivity period during which they can search for the ideal candidate without competition from other agencies. In a similar line, 'no cure no pay' contract conditions force agents to close the deal as soon as possible if they want to forego the danger of their time and effort invested going to waste because they were outrun by some other agency. Institutional theory argues that organizations are more likely to comply to a constituent's calls when the consequences of non-compliance are highly punitive and strictly enforced. In the case of a client hiring a search agent to look for a particular type of candidate, this thus implies that search agents' concern for speed will be stronger under less agent-friendly contract conditions, i.e., the more they are competing on an open market (without an exclusivity period) for finding a suitable candidate while expecting to get paid only after the successful completion of the project (no cure no pay). Thus we propose:

Hypothesis 2a: The speedier professional search agents perceive a specific recruitment procedure to lead to a suitable candidate, the more likely they are to use this procedure.

Hypothesis 2b: This effect will be moderated by the type of contract that search agents use with their clients. The more their contracts tend towards a client-friendly solution based contract ('no cure no pay'), the more likely they will be to pay attention to the speed with which a specific recruitment procedure will lead to a suitable candidate.

Longevity of solution found. Finding candidates fast is likely to foster short-term success, but does not guarantee long-term commitment and thus long-term business from search agents' clients. A lack of trust is a major issue in the professional search business (Finlay & Coverdill, 2002) and client organizations cannot always be sure that the solution proposed by search agents is one that not only looks good at first sight but that also promises to be sustainable. After all, one of the search agents' tasks in the process of grooming candidates and organizations to match up is to make both sides see all the possible benefits of such a match and to essentially 'fall in love' with one another (Finlay & Coverdill, 2002). Entering the new job, many new employees then experience a thoroughly positive period of new possibilities and the satisfaction of previously unsupported needs. Yet, such honeymoon usually lasts only a few months before reality kicks in and the new employees may not only be less euphoric about their new job but may actually find themselves in a hangover type of disappointment (Boswell, Boudreau, & Tichy, 2005). This really is the time then when both candidates and client organizations can judge whether the match arranged by the search agent is going to be sustainable.

For search agents, providing client organizations with truly suitable candidates who will commit to the client also beyond the few-month-honeymoon phase is a major source of credibility and reputation as a reliable broker of suitable candidates. Yet again, the importance of such reputation will be particularly relevant under two conditions: first, if agents see a general opportunity for follow-up work coming from the client in question and second, if the candidate market is particularly tight, thus enhancing the likelihood of candidates on the market to switch positions again and again. With the task of matching client organizations and candidates, these two groups are actually the constituents or stakeholders that professional search agents will be most dependent upon, and both of these stakeholders can exhibit control over the search agents by threatening to turn away from them. The consequences of constituents exercising their control, however, also depend on the context (Oliver, 1991). Thus, the possible loss of credibility and ensuing reprimands by client organizations on search agents for placing a poor or uncommitted candidate with the organization will be greater if the search agent had any reason to hope for follow-up contracts from this client organization. One-time-only clients, however, hold relatively little power over search agents in this regard (possibly because they look small and insignificant or because they are known to show no loyalty to any search agent, anyway; Finlay & Coverdill, 2002).

In a similar line, concerns for sustainable solutions will be particularly high in a labor market in which suitable candidates are scarce and in which thus candidates may be approached regularly by different search agents and may regularly face the choice between multiple employers, easily being recruited away from their new spot to other organizations. Consequently, we propose that:

Hypothesis 3a: The more professional search agents perceive a specific recruitment procedure to produce candidate that will stay with the organization for a certain time, the more likely they are to use this procedure.

Hypothesis 3b: This effect will be moderated by the perceived likelihood of client organizations to offer follow-up work to the search agents. The more promising clients appear for follow-up contracts, the more attention search agents will pay to finding candidates that do stay with the client organization for a certain amount of time.

Hypothesis 3c: Effect 3a will also be moderated by the perceived tightness of the candidate market. The fewer suitable candidates an agents perceives to be available for a possible position, the more attention the search agent will pay to finding candidates that do stay with the client organization for a certain amount of time.

Social Fitness

Touching upon the two main constituents or stakeholders whom search agents need to satisfy during their work also points at the other major cause that organizations need to satisfy beside economic fitness, namely their social fitness (Oliver, 1991). According to both resource dependency and institutional theory, organizations strive to prove their social worthiness or legitimacy (Oliver, 1991) and "the general theme of the institutional perspective is that an organization's survival requires it to conform to social norms of acceptable behavior" (Covaleski & Dirsmith, 1988, p. 563). Such conformity is by no means self-sufficient but rather comes with numerous advantages such as increased prestige, stability, legitimacy, social support, internal and external commitment, access to resources, attraction of personnel, fit into administrative categories, acceptance in professions, and invulnerability to questioning (DiMaggio, 1988; DiMaggio & Powell, 1983; Meyer & Rowan, 1977, 1983; Meyer & Scott, 1983; Scott, 1983; Zucker, 1988). In the case of third-party recruiters, much of this boils down to the simple question: Will recruiters be accepted as required and respectable brokers by both client organizations and candidates?

Candidate Reactions

Besides eliciting a speedy response from potential candidates, a successful matchmaking also requires that this response be positive. Cold calling complete strangers who may be happily settled in their current position, for example, with the aim of turning them into potential candidates interested in greener pastures elsewhere and making them believe that the recruitment agent truly has just such perfect opportunity at hand, is an admittedly awkward and difficult task (Finlay & Coverdill, 2002).

Candidate reactions are a long-standing concern in the recruitment and selection literature, as procedures that decrease an organization's attractiveness result in significant losses in financial utility when qualified candidates refuse job offers (Murphy, 1986) or even withdraw before receiving an offer (e.g., French, 1987; Murphy & Davidshofer, 1988). In an extreme case, candidates may perceive recruitment procedures to be unethical or immoral, thus refusing further collaboration or possibly even decrying the recruitment agency to their current or later employer.

At the same time, the consequences of such effect will again differ and will likely depend on all three moderators that we have discussed so far: the type of contract between agent and client, the perceived tightness of the candidate market, and the perceived likelihood of client organizations to offer follow-up work. The type of contract will likely function as a moderator of the impact of candidate reactions, with negative candidate reactions being perceived as particularly severe under 'no cure no pay' conditions. While more recruiter-friendly contracts may allow agents to still get paid if they do not manage to recruit a suitable candidate, 'no cure no pay' conditions raise the likelihood that a disgruntled candidate will translate into the search agent's time and energy into the search being wasted. The same applies to tight candidate market when one disgruntled candidate quickly translates into a failure to find a suitable candidate for the open position. Finally, negative candidate reactions will also become particularly averse in situations in which the agent hopes to establish a long-term relationship with the client. Losing suitable candidates for open positions undermines the agent's credibility as someone who can find the right people for the jobs at hand. Not losing a candidate yet placing a candidate who feels sceptical about the search agent's approach with the client may have negative consequences as well, given that the candidate will likely sooner or later share his or her negative perceptions of the recruitment process with the hiring manager, thus again undermining the agent's credibility and perceived trustworthiness. Thus we propose that:

Hypothesis 4a: The more negative professional search agents expect the candidates' reaction to a specific recruitment procedure to be, the less likely they are to use this procedure.

Hypothesis 4b: Effect 4a will be moderated by the type of contract that search agents use with their clients. The more their contracts tend towards a client-friendly solution based contract ('no cure no pay'), the more attention search agents will pay to likely candidate reactions.

Hypothesis 4c: Effect 4a will also be moderated by the perceived likelihood of client organizations to offer follow-up work to the search agents. The more promising clients appear for follow-up contracts, the more attention search agents will pay to candidate reactions.

Hypothesis 4d: Effect 4a will also be moderated by the perceived tightness of the candidate market. The fewer suitable candidates an agents perceives to be available for a possible position, the more attention search agents will pay to candidate reactions.

Client Reactions

The same importance can be placed on the clients who decide whether and when to pay search agents for their services. Therefore, agents have a strong incentive to appear able to solve client's needs. In other words, agents gain social legitimacy by using recruitment procedures that clients, too, perceive to be effective in finding suitable candidates. Again, however, the strength of this effect is likely to depend on the agent's dependency on this client in terms of the client's ability to provide the agent with further business. Thus, we assume that

Hypothesis 5a: The more professional search agents believe clients to perceive a certain recruitment strategy to be effective, the more likely they are to use this procedure.

Hypothesis 5b: Effect 5a will be moderated by the perceived likelihood of client organizations to offer follow-up work to the search agents. The more promising clients appear for follow-up contracts, the more attention search agents will pay to client reactions.

Context: Dissemination in the Field

Finally, Oliver (1991) argued that several factors of an organization's context also influence an organization's reaction, most prominently among them the diffusion of comparable procedures in the field and the uncertainty of the organization's context. The diffusion of a comparable recruitment procedure in the field is likely to influence the perceived legitimacy and thus adoption of this procedure. Referring back to the mimetic view of organizational conformity (DiMaggio & Powell, 1983; Tolbert & Zucker, 1983), Oliver (1991) hypothesized that the extent to which an institutional expectation or practice has already diffused or spread voluntarily through an organizational field will tend to predict the likelihood of conformity. When institutional rules or norms are broadly diffused and supported, their social validity is largely unquestioned. They reflect taken-for-granted understandings of what constitutes legitimate or rational behavior. Organizations may even conform because it does not occur to them to do otherwise. This is less likely to occur when specific values or practices are not broadly supported.

Again, however, this effect is likely to be moderated by the context in which recruitment services do their business. More specifically, institutional theory (DiMaggio & Powell, 1983) proposes that in the context of uncertainty, i.e., when "future states of the world cannot be anticipated and accurately predicted" (Pfeffer & Salancik, 1978), organizations will strive to re-establish a sense of control and predictability, ending up in imitating one another. Thus, we assume that

Hypothesis 6a: The more professional search agents believe other agents to use a certain recruitment strategy, the more likely they are to use this procedure.

Hypothesis 6b: Effect 6a will be moderated by the perceived uncertainty of the recruitment business. The less certain the recruitment business appears to be, the more professional search agents will imitate the recruitment procedures used by other recruiters.

Method

Participants and Procedures

This study was conducted in cooperation with a Dutch branch organisation for search and selection practices. Four hundred and fifty representatives from professional search and selection companies received an email inviting them to participate in a benchmarking study on recruitment practices in the Netherlands and accompanied with instructions about and a direct link to the online survey. In order to improve the response rate, a second email was sent by the chairman of the branch organization who also requested respondents to fill in the survey. Two weeks later, a reminder was send to all potential participants. Another two weeks later we started calling members of the branch organization, asking them if they had received our emails and politely requesting them to fill out the questionnaire.

Of the 450 people contacted, 237 opened the link to the online survey. From these 237 participants, 149 responded at least 50% of the questionnaire and 115 completed it to the end. The gender distribution was 45% female versus 55% male. Most participants held a masters' (53%) or a bachelors' degree (39%), with remaining levels of qualifications ranging from high school diploma (2%) to post master degree (2%). Their average age was 38 (SD=10.03; spanning from 22 to 67 years) and participants had been working in their current position for an average of 7.5 years (SD=6.05). For all participants the recruitment of candidates was part of their jobs. In average, recruiters themselves had placed 17 candidates within the last year, while the mean and median for their organization were 153 and 62 candidates, respectively. Recruiters' clients span across numerous industries, most prominent being clients in information technology (28%), banking (26%), electronics (23%), logistics (23%), retail (19%), food and consumer goods (19%), pharmaceutical and medical products as well as healthcare (both 14%) and real estate (14%). Most prominently searched were candidates in middle (90%) and top management (78%), followed by specialists (59%), lower management (44%), and technical staff(30%), and finally administrative staff(23%), starters (21%) and management trainees (10%).

Measures

The survey was developed on the basis of the theoretical model as well as earlier questionnaires aimed at the related literature of personnel selection (Glode, 2002; König et al., 2010). To test for clarity and accuracy of the questions, 7 graduate students of personnel psychology and 3 professional recruitment consultants pre-tested the questionnaire, leading to a slight adaptation of some questions and an elimination of items that consultants had deemed to be too repetitive and thus likely to lower recruiter participation in the study.

The survey consisted of five parts, (1) addressing respondents' own position and organisation, (2) the search procedures used (including the dependent variables), (3) the type of contract on which respondents conducted their business and their evaluation of the current employment market (i.e., the moderators proposed), (4) detailed questions about selected search procedures (i.e., the predictors of the proposed main effects), and finally (5) demographic information. Asking participants first about the dependent variables (among others) before asking them about the moderators and then the main predictors served to reduce the likelihood of consistency biases that might otherwise account for the findings reported. On average, respondents needed 24 minutes to complete the questionnaire.

Section 1: Descriptives. This section served to gain an overview of the search and selection industry. Questions addressed position, type of organisation, one's own tasks (e.g., searching for suitable candidates, assessment of candidates, acquisition of new clients) and the type of branch the consultant was active in (e.g., healthcare, information technology, etc.).

Section 2: Dependent Variables. In order to assess the use of different search procedures, we asked participants: While searching for candidates over the past 12 months, how often have you used the following search procedures? (please indicate the frequency for every procedure), followed by a detailed list of 25 procedures subdivided into five categories (see Table 1). The categories were: "Advertising and job sites", "Online business networks", "Direct search", "Desk search" (such as newspapers, magazines, yearbooks), and "Alternative search" (such as job fairs or campus recruitment). Frequencies were indicated on a five point Likert scale with options labelled as never, seldom, sometimes, often, and always. We also provided a free field in which recruiters were able to write down additional procedures. While this information served to gain a comprehensive overview of the procedures actually used in the field, hypotheses were tested on recruiters' use of (a) advertisements, (b) online networks, and (c) direct search, as these, so we were told by the recruitment consultants serving as pre-test participants, reflected the search procedures most representative of the field (confirmed by our findings, see Table 1), and in order to keep the survey short enough to ensure a sufficient response rate.

Advertisements. Recruiters indicated their use of online advertisements and print advertisements, in addition to six specific advertisement sites, comprising of international sites such as Monsterboard and Stepstone but also of national sites such as the Nationale Vacaturebank and Intermediair. Cronbach's alpha across the eight items was .81.

Online networks. Recruiters indicated their use of six online networks, such as Linkedin, Xing, but also the Dutch network Hyves and Google. Cronbach's alpha across the six items was .74.

Direct search. Recruiters indicated their use of direct search methods by indicating how often they had searched their own database, contacted old contacts /colleagues or private networks or had directly searched for suitable candidates in comparable positions at the clients' competitors. Cronbach's alpha across the six items was .60.

Section 3: Moderators. The third section of the questionnaire contained questions concerning the payment structure underlying the contract between the recruiter and the client organization as well as their perceptions of the market they are dealing with, both of which are proposed to moderate the proposed main effects.

Contract structure. Search agents indicated on a scale from 1 (never) to 5 (always), how often their contracts included the eight different contract structures 'no cure no pay', retainer, exclusivity period, fee after client signs the contract, fee after candidate shortlist, fee after candidate signs contract with organization, fee at start of candidate in client's office, fee after guarantee period. All these items loaded onto a single factor, indicating that all of them described the degree to which a contract tended to be agent-friendly and effort based (e.g., retainer, exclusivity period) or tended to be solution based and thus client-friendly (e.g., no cure no pay, guarantee period). After reverse coding those items that indicated that search agents received any benefits before having successfully completed the search (i.e., the agent- friendly items), Cronbach's alpha was .87.

Likelihood of follow-up work. Search agents indicated on three items to what degree their clients usually had the potential to provide them with new/follow-up assignments in the future or to recommend their services to other potential clients [e.g., To what degree do clients have the potential to give you new assignments in the future?, rated on a scale from 1 (never) to 5 (always)]. Cronbach's alpha was .61.

Candidate market. Search agents indicated their perception of the market with three items asking them the number of reactions in total and particularly promising reactions to the last vacant position they had worked on, and also to indicate in more general terms their perception of the candidate market [e.g., To what degree do you agree there are more candidates than job openings available on the market?, rated on a scale from 1 (never) to 5 (always)]. Cronbach's alpha was .75.

Environmental uncertainty. Search agents indicated their perception of the state of the recruitment business with the two items it is a hard time for the search and selection industry at the moment and there is a good position in the market for search firms right now (reverse coded), rated from 1 (highly disagree) to 5 (highly agree). Cronbach's alpha was .59.

Section 4: Predictors. We asked recruitment agents to evaluate each of the three recruitment approaches (advertisement, online networks, direct search) on all of the different predictors proposed above. Given that this required us to ask the following questions three times (first on advertisement, then on online networks, and finally on direct search) and given that we needed to keep the questionnaire short and respondents motivated rather than bogged down by repetition, we only asked one or two questions per predictor proposed, always answered on a five point Likert scale ranging from 1 (not likely at all) to 5 (very likely).

Cost for procedure. Perceived cost was measured by asking: How likely is ... (search procedure) to be expensive? Time Investments was measured by asking: How likely is ... (search procedure) to require high investments of time? While we had originally planned to combine these two items into one measure, the low correlation (.29) between them suggests that they be better treated separately. Response speed of candidates was measured by asking: How likely is ... (search procedure) to lead to a quick response of possible candidates? Candidate stay: The likelihood that candidates will stay with the client organization for a minimum amount of time was measured with two items, namely How likely is ... (search procedure) to lead to candidates who will be pleased with their new job? and How likely is ... (search procedure) to lead to candidates who will stay with the client's organisation for at least 6 months? Candidates' likely reactions to the recruitment effort were measured with the reverse-coded item How likely is ... (search procedure) to annoy a candidate you contacted? Perceived effectiveness by client was measured with How likely is ... (search procedure) to be perceived as effective by the client? Diffusion in the field was measured by asking How likely is ... (search procedure) to be used by other search companies in the field?

Results

Preliminary Analyses

Search procedures. Figure 2 shows an overview of the frequency with which participants used the 25 search procedures included in the survey. Most prominent by far were recruiters' own database from which to draw up candidates (89%) as well as the use of online advertisement (78%) and personal contacts (70%), followed by a number of online job and business network sites. Recruiters also acknowledged to sometimes searching their clients' competitor organizations for suitable candidates.

Contract. When addressing the contract structure under which professional search agents usually search, results showed that recruiters often work with an exclusivity period (74%), allowing them to be the sole provider of a suitable candidate within a given time frame, rather than competing with other professional search agents already from the start. Regularly, recruiters are paid a certain fee at the onset of the task (76%) and/or work from a retainer fee (47%). Fees at the time that the search agent presents the organization with a shortlist of potential candidates is somewhat common (48%), while most agents collect the final parts of their fees once the candidate has signed the contract with the client organization (84%). After that, only relatively few recruiters (28%) report receiving a fee once the candidate has actually started working at the client organization or after a certain guarantee period (21%). Different to the data reported by Finlay and Coverdill (2002), only 41% of the participating recruitment agents worked on a 'no cure no pay' basis.

Evaluations of recruitment procedures. Comparing the three recruitment approaches studied in more detail (Table 1), results reveal rather diverging patterns. Recruiters evaluated direct search as the most time-consuming and least common recruitment-procedure available, yet also the procedure that was most valued by clients and that lead to the highest likelihood of recruiting candidates who would stay with the client organization for a reasonable length of time. The use of online networks struck recruiters as particularly cheap, yet also as the procedure least valued by clients and most likely to lead to candidates who show only a short-term commitment to the recruiters' clients. Finally, the use of advertisements was judged to be the least time-consuming but also slowest procedure to elicit candidate responses. Being perceived as the most common in the field, this procedure was also judged as least likely to elicit negative candidate reactions. Overall, these results thus suggest sufficient variance between recruitment procedures on the different predictors proposed to test for the effects that each of these predictors may have on the actual usage of different recruitment procedures.

Testing of Hypotheses

Given that each search agent evaluated three different recruitment procedures, data are nested within individuals, calling for a hierarchical analytical approach. We first used multilevel modelling (SPSS mixed models) to check how much variance in the dependent variable "usage of search procedures" was due to differences between participants. These analyses take into account that measurements are repeated within participants and are thus not independent of one other by allowing a parsimonious examination of within and between person variance (Hox, 2002). We started with an intercept-only model or null model, a model without any predictors but with searching procedure as the repeated factor, as a reference model. We estimated the intraclass correlation (ICC) ρ, which reflects the ratio of variance between persons to the variance within persons. The resulting ICC indicated that only relatively little variance can be explained by individual differences (ρ <.01). This could be explained by our focus on three rather different recruitment approaches, due to which the within-person variance is higher than the between-person variance.

Against the null model we tested a model with all the proposed main effects/predictors in order to test for their relationship with usage of the different search procedures. This model provides information on the average contribution of each predictor across participants. We computed the explained variance accounted for by the predictors as the total variance observed in the null model minus the variance not attributable to the predictors, divided by the total variance. Averaged across participants, our proposed predictors accounted for almost 26% of the variance in usage of search procedures. Subsequently, we added our proposed moderators to the model, accounting for another 16% in variance. Below we will summarize the findings per Hypothesis (see also Table 3).

Contrary to our Hypothesis 1a, participants' perception of how cost intensive and how time intensive a specific procedure was, was positively related to their likelihood to use this procedure (b=.17; p<.01; b=.20; p<.01). However, as proposed in Hypothesis 1b, the effect of time intensity was moderated by the type of contract that participants use with their clients, with recruiters being less willing to take up such extra expenses when they are unsure of being reimbursed for them in the end (b=-.15; p<.05; see Figure 3).

Furthermore, findings supported our Hypothesis 2a, the speedier participants perceive a specific procedure to lead to a suitable candidate, the more likely they are to use this procedure (b=.17; p<.01). This effect, contrary to expectations, was not moderated by the type of contract that recruiters used with their clients. Hypothesis 3a was also supported: participants' expectations of whether candidates will stay with the organization for a certain time was related to their likelihood to use this procedure (b=.20; p<.01), an effect moderated by the perceived tightness of the labor market (b=.20; p<.05; see Figure 4), thus supporting Hypothesis 3c.

As proposed in Hypothesis 4a, participants were less likely to use a procedure if they expected the candidates to react more negative to the procedure (b=-.27; p<.01). Effect 4a was furthermore marginally moderated by the type of contract, tentatively confirming Hypothesis 4b (b=.12; p<.10; Figure 5) and was fully moderated by the perceived likelihood of client organizations to offer follow-up work confirming Hypothesis 4c (b=.16; p<.01; Figure 6). The more their contracts tend to solution based and the more promising the client appears for follow-up work the more participants pay attention to candidate reactions.

No support was found for Hypothesis 4d, the moderation of the effect by the perceived tightness of the labor market, however. As proposed in Hypothesis 5a, participants' expectations about their clients' perceptions of the effectiveness of a certain recruitment strategy were indeed positively related to their likelihood to use a specific procedure (b=.26; p<.01). Yet, the perceived likelihood of client organizations to offer follow-up work to the participants did not moderate this relation. Hypothesis 5b was thus not supported. Finally, as proposed in Hypothesis 6a, as participants believe other agents to use a certain recruitment strategy, the more likely they are to use this procedure (b=.24; p<.01) and this is especially true, contrary to Hypothesis 6b, as the recruitment business appears as more certain (b=-.12; p<.05; see Figure 7).

Discussion

Professional search agents play a dominant role in matching qualified candidates with management and professional positions, i.e., in recruiting the people who shape technological advancements or take managerial responsibility for the direction and thus the future of organizations. Despite this dominant role, professional search agents and the reasons shaping their decisions to rely on the one or the other recruitment approach have hardly been studied. The current study tried to fill this gap by asking search agents about the recruitment procedures they used and their evaluation of several of these search procedures.

Overall, search agents indicated substantial variance in how they search for candidates with most procedures listed being used often or always by some and never or seldom by others (with the exception of the 'alternative search' procedures and some technical search options that may originally have been designed for other purposes or audiences). An exception is the use of the own database, something that agents often claim to be a major capital in terms of access to specialized pools of candidates. At the same time, the relatively high frequency of sieving through a competitor's organization to find an adequate candidate also stands well in line with earlier writing about headhunters and with the bad reputation that they might have gained.

When it comes to predicting the search procedures used, results supported the importance that search agents place on finding both fast (H2a) and sustainable (H3a) solutions, while also satisfying their primary stakeholders, namely client organizations (H5a) and potential candidates (H4a). Further, as suggested by institutional theory, search agents tended to use procedures more, the more they believed the procedures to be common in the search industry overall (H6a). Different from expectations (H1a), however, search agents did not shy away from expensive or time-intensive procedures. A possible reason for this may be that the majority of search agents were not only paid for the outcome they produced, but also for the effect that they invested, with about ¾ of agents receiving an initial payment at the start of the contract and about half of them working from a retainer. Another reason may be that search agents actually perceived more expensive and time-consuming search procedures to be more likely to be embraced by clients and to produce candidates who will prove to be a sustainable addition to the client organization. At the same time, however, results still confirmed the proposition stemming from institutional theory that agents' readiness to invest such effort will depend on the security they have that such effort will actually pay off. That is, agents particularly reported using time-intensive procedures when they were working under agent-friendly contractual conditions in that their income was, at least in parts, secured, they were paid on an effort, rather than results basis, and were working from a retainer. As expected, agents were, however, less eager to invest into time-intensive search methods when the outcome of the search for their own financial benefits was still uncertain (supporting H1b).

While we found no support for the proposed moderating influence of contract structure on the importance of speed in finding suitable candidates (H2b), contract structure also influenced the importance that agents placed on the positive reactions of one of their core stakeholders, namely candidates (H4b). More specifically, while recruiters were generally rather concerned about not annoying potential candidates with unsolicited contacts and requests, these concerns were particularly relevant for agents who were contractually dependent on leaving candidates with a good impression if they wanted to make an income.

In a similar line is the finding that preventing negative candidate evaluations proved particularly relevant when search agents were expecting follow-up work from the same client (supporting H3b), even though this proposed moderator showed no significant impact either on the relevance of finding a sustainable solution. Results were opposite, in comparison, for the perceived tightness of the candidate market, with such tightness moving agents to pay particular attention to finding a sustainable solution (supporting H3c), not, however, to pay more attention to candidates' immediate reactions to the recruitment approach. Finally, the perceived uncertainty of the business environment indeed moderated the impact of the dissemination of a particular recruitment procedure on search agents relying on this procedure as well. The shape of this moderation, however, was different that we had expected. Rather than agents showing more mimicry and imitation during difficult times for the industry, they showed more willingness to also rely on less common recruitment procedures under these conditions. While this decision contradicts propositions derived from institutional theory, however, it does make sense from a business perspective, since it implies that when perceiving the market as non-favourable, search agents start considering new and alternative approaches to getting their work done and possibly finding a niche that they feel less preoccupied by their competitors.

Implications for Practice and Conceptual Contributions

The current paper makes a number of practical and conceptual contributions. Practically, the data reported here provide the first, we think, overview of recruitment practices currently used, and while some specificities might change (e.g., some social network sites may gain in popularity while others may go down), such overview may serve as a benchmark and as inspiration for further research examining the business of professional search firms.

Conceptually, the study contributes to the thin knowledge that we have about professional recruiters. While considerable research has studied the case of personnel selection, the literature on recruitment is still relatively scarce, and the literature on the experts of recruitment, who try to win particularly promising candidates for top level positions, is even scarcer. The current paper provides a first overview over the recruitment procedures used and the reasons why recruiters decide on one or the other procedure. When doing this, we largely relied on institutional theory to explain the proposed effects; and indeed, results largely supported our assumptions and with this the applicability of institutional theory also to the case of personnel recruitment. Besides offering another case on the benefits of institutional theory, these results are also relevant for research related to personnel psychology and human resource management. They remind us, after all, that organizational decisions happen within a context of diverging pressures and stakeholders, and also personnel decisions are not exempt from this. Rather, 'micro' research originating from disciplines such as psychology may well benefit from eyeing over to more 'macro' research found in organizational theory (and vice versa, likely). This will also be particularly relevant in discussions about the scientist practitioner gap in personnel management, for example. Thus, an organizational theory perspective on the science practitioner gap in personnel selection (Klehe, 2004) revealed that rather than being upset about the lack of compliance among practitioners with the selection procedures recommended by science, science may benefit from considering the context of personnel selection -thus realizing that (a) the voice of science is small and harmless (in terms of institutional control mechanisms and reprimands in case of non-compliance) in organizational decision making and that (b) organizations have to respond to a wide chorus of different requirements, some of which have hardly been sufficiently studied by the scientific community.

The latter is definitely also true for the case of recruitment. While we know, for example, that cognitive ability tests usually do a better job at identifying suitable candidates than do unstructured interviews, no such information is available on different recruitment procedures, despite good recruitment being the prime pre-requirement for even having suitable candidates to select from. In this regard, the current study had to fully rely on recruiters' own perceptions of different procedures, yet we do not know how accurate these perceptions truly are. Given from what we know about stubborn decisions in personnel selection (Highhouse, 2008), it would not surprise if also the recruitment industry was pervaded by erroneous myths.

Limitations and Directions for Future Research

To our knowledge, this was the first study addressing the procedures used by professional search firms to find suitable candidates for their clients. Thus, we largely had to develop measures ourselves, rather than relying on pre-existing measures. This, of course, calls to question the construct related validity of the measures employed. We tried to evade this problem by developing items that were as straight forward as we could phrase them and by pretesting the measures with the help of search agents working in the field. As it turned out, agents agreed on the purpose of our items but recommended that we shorten the questionnaire of originally three items per construct to one if we wanted to secure a meaningful response rate. As it turned out during the actual data collection, search agents were indeed rather hesitant to report about their business and their opinion on different search procedures and did not always react positively to being asked the same questions several times in order to gain their evaluations of different recruitment procedures. Possibly, this really is one of the primary reasons why past research on professional search agents or headhunters was primarily been qualitative rather than quantitative in nature. Researchers such as Finlay and Coverdill (2002) invested a considerable amount of time into getting to know their participants and into winning their trust, an initiative that is hardly available to quantitative researchers.

On the other hand, quantitative research offers the benefits of anonymity. This is particularly relevant in the case of single-source data collected at a single point in time, which may otherwise invite socially desirable responding and common method variance. Several methodological decisions and empirical findings speak against such danger, however. Methodologically, we (a) guaranteed respondents' anonymity to the questionnaire, and (b) tried to phrase questions as straight forward and objective as possible. We (c) enquired about the more socially desirable dependent variable before asking participants about the independent variables, and (d) separated the two sections by having another section between them asking participants about the variables we assumed to moderate our effects (as well as a number of variables of interest to the branch organisation with whom we were collaborating on this project). Empirically, results suggest that search agents responded candidly to our questions. After all, they admitted to not taking a number of initiatives that might otherwise be perceived as promising for finding suitable candidates, yet also largely admitted to raiding clients' competitors for potential candidates. Further, the overall results found can hardly be explained by common method variance and/or social desirability. For one, correlations between distinct variables were generally rather low and each of the proposed main-effects turned out to be significant overall. For the other, multiplicative terms as used in calculating interactions are usually too complex to expect or to mimic by respondents, and yet, the data still supported the majority of proposed interactions.

Conclusion

The current study presented the first test of professional search firms and the recruitment procedures that they rely upon. Results confirmed that similar to personnel selection (Klehe, 2004; König et al., 2008), much of practitioners' recruitment decisions could be explained by institutional theory, highlighting the complexities that processional recruiters encounter when trying to stay in business while matchmaking clients and candidates

References

References

Boswell, W.R., Boudreau, J.W., & Tichy, J. (2005). The relationship between employee job change and job satisfaction: The honeymoon hangover effect. Journal of Applied Psychology, 90, 882-892.

Bowe, J., Bowe, M., & Streeter, S. (Eds.). (2000). GiG Americans talk about their jobs. New York: Three Rivers Press.

Caeldries, F. (1996). The institutional embeddedness of strategy: Predation through legislation (or, I'll see you in court). In J.A.C. Baum & J.E. Dutton (Eds.), Advances in Strategic Management (Vol. 13, pp. 215-246). New York: JAI Press.

Collins, J. (2007). The interactive effects of recruitment practices and product awareness on job seekers' employer knowledge and applicant behaviour. Journal of Applied Psychology, 92, 180-190.

Collins, C.J., & Hahn, J. (2004) Exploring applicant pool quantity and quality: The effects of early recruitment practice strategies, corporate advertising and firm reputation. Personnel - Psychology, 57, 685-717.

Covaleski, M.A., & Dirsmith, M.W. (1988). An institutional perspective on the rise, social transformation, and fall of a university budget category. Administrative Science Quarterly, 33, 562-587.

Coverdill, J. E., & Finlay, W. (in press). Headhunters. In U.-C. Klehe & E. Hooft(Eds.) Oxford handbook of job loss and job search. New York, NY: Oxford University Press.

DiMaggio, P.J. (1988). Interest and agency in institutional theory. In L.G. Zucker (Ed.), Institutional patterns and organizations: Culture and environment (pp. 3-21). Cambridge, MA: Ballinger.

DiMaggio, P.J., & Powell, W.W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 48, 147-160.

Finlay, W., & Coverdill, J.E. (1999). The search game: Organizational conflicts and the use of headhunters. Sociological Quarterly, 40, 11-30.

Finlay, W., & Coverdill, J.E. (2000). Risk, opportunism, and structural holes - How headhunters manage clients and earn fees. Work and Occupations, 27, 377-405.

Finlay, W., & Coverdill, J.E. (2002). Headhunters: Matchmaking in the labor market. Ithaca, NY: Cornell University Press.

Fish, A., & Macklin, R. (2003). Perceptions of executive search and advertised recruitment attributes and service quality. Personnel Review, 33, 30-54.

French, W.L. (1987). The personnel management process. Boston: Houghton Mifflin.

Glode, J. (2002). Expanding the justice dilemma: A comparison of stakeholder reactions to personnel selection tools (Unpublished doctoral dissertation). Portland State University, Portland.

Goodstein, J.D. (1994). Institutional pressures and strategic responsiveness: Employer involvement in work-family issues. Academy of Management Journal, 37, 350-382.

Heraty, N., & Morley, M. (1998). In search of good fit: Policy and practice in recruitment and selection in Ireland. Journal of Management Development, 17, 662-685.

Highhouse, S. (2008). Stubborn reliance on intuition and subjectivity in employee selection. Industrial and Organizational Psychology, 1, 333-342.

Hox, J.J. (2002). Multilevel analysis: Techniques and applications. Mahwah, NJ: Erlbaum.

Ingram, P., & Simons, T. (1995). Institutional and resource dependence determinants of responsiveness to work-family issues. Academy of Management Journal, 38, 1466-1482.

Khurana, R. (2002). Market triads: A theoretical and empirical analysis of market intermediation. Journal for the Theory of Social Behaviour, 32, 239-262.

Khurana, R. (2004). Headhunters: Matchmaking in the labor market. Journal of Economic Literature, 42, 208-209.

Klehe, U.C. (2004). Choosing how to choose: Institutional pressures affecting the adoption of personnel selection procedures. International Journal of Selection and Assessment, 12, 327-342.

König, C.J., Klehe, U.C., Berchtold, M., & Kleinmann, M. (2010). Reasons for being selective when choosing personnel selection procedures. International Journal of Selection and Assessment, 18, 17-27.

Lavelle, L. (2003, September 29). After the jobless recovery, the war for talent. Business Week, pp. 92.

Maher, K. (2003, January 14). Corporations cut middlemen and do their own recruiting. The Wall Street Journal, pp. B10.

Meyer, J.W., & Rowan, B. (1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83, 340-363.

Meyer, J.W., & Rowan, B. (1983). The structure of educational organizations. In J.W. Meyer & W.R. Scott (Eds.), Organizational environments: Ritual and rationality (pp. 71-97). Beverly Hills, CA: Sage.

Meyer, J.W., & Scott, W.R. (1983). Centralization and the legitimacy problems of local government. In J.W. Meyer & W.R. Scott (Eds.), Organizational environments: Ritual and rationality (pp. 199-215). Beverly Hills, CA: Sage.

Murphy, K.R. (1986). When your top choice turns you down: Effect of rejected job offers on the utility of selection tests. Psychological Bulletin, 99, 133-138.

Murphy, K.R., & Davidshofer, C.O. (1988). Psychological testing. Engelwood Cliffs, NJ: Prentice Hall.

Oliver, C. (1991). Strategic responses to institutional processes. The Academy of Management Review, 16, 145-179.

Pfeffer, J. (1981). Power in organizations. Boston, MA: Pitman.

Pfeffer, J., & Salancik, G.R. (1978). The external control of organisations. New York: Harper & Row.

Schneider, B. (1987). The people make the place. Personnel Psychology, 40, 437-453.

Scott, W.R. (1983). Health care organizations in the 1980s: The convergence of public and professional control systems. In J.W. Meyer & W.R. Scott (Eds.), Organizational environments: Ritual and rationality (pp. 99-113). Beverly Hills, CA: Sage.

Tolbert, P.S., & Zucker, L.G. (1983). Institutional sources of change in the formal structure of organizations: The diffusion of civil service reform, 1880-1935. Administrative Science Quarterly, 28, 22-39.

Van Iddekinge, C.H., & Ployhart, R.E. (2008). Developments in the criterion-related validation of selection procedures: A critical review and recommendations for practice. Personnel Psychology, 61, 871-925.

Zucker, L.G. (1977). The role of institutionalization in cultural persistence. American Sociology Review, 42, 726-743.

Zucker, L.G. (1988). Where do institutional patterns come from? Organizations as actors in social systems. In L.G. Zucker (Ed.), Institutional patterns and organizations: Culture and environment (pp. 23-49). Cambridge, MA: Ballinger.

Received: August 30, 2012